| | 290 | === ATA drive is failing self-tests, but SMART health status is 'PASSED'. What's going on? === |
| | 291 | |
| | 292 | If your ATA drive supports self-tests, you should run them on a regular basis, for example one per week: |
| | 293 | {{{ |
| | 294 | smartctl -t long /dev/hd? |
| | 295 | }}} |
| | 296 | |
| | 297 | After the test has completed, you should examine the results with: |
| | 298 | {{{ |
| | 299 | smartctl -l selftest /dev/hd? |
| | 300 | }}} |
| | 301 | |
| | 302 | If the drive fails a self-test, but still has '{{{PASSED}}}' SMART health status, this usually means that there is a corrupted (uncorrectable=UNC) sector on the disk. This means that the ECC data stored at that sector is not consistent with the user data stored at that sector, and an attempt to read the sector fails with a UNC error. This can be a one-time transient effect: a sudden power failure while the disk was writing to the sector corrupted the ECC code or data, but the sector <em>could</em> correctly store new data. Or it can be a permanent effect: the magnetic media has been damaged by a bit of dust, and the sector could ''not'' correctly store new data. |
| | 303 | |
| | 304 | If the disk can read the sector of data a single time, and the damage is permanent, not transient, then the disk firmware will mark the sector as 'bad' and allocate a spare sector to replace it. But if the disk can't read the sector even once, then it won't reallocate the sector, in hopes of being able, at some time in the future, to read the data from it. '''A write to an unreadable (corrupted) sector will fix the problem.''' If the damage is transient, then new consistent data will be written to the sector. |
| | 305 | If the damange is permanent, then the write will force sector reallocation. Please see [http://smartmontools.sourceforge.net/badblockhowto.html Bad block HOWTO] for instructions about how to force this sector to reallocate (Linux only). |
| | 306 | |
| | 307 | The disk still has passing health status because the firmware has not found other signs of trouble, such as a failing servo. |
| | 308 | |
| | 309 | Such disks can often be repaired by using the disk manufaturer's 'disk evaluation and repair' utility. Beware: this may force reallocation of the lost sector and thus corrupt or destroy any file system on the disk. See [http://smartmontools.sourceforge.net/badblockhowto.html Bad block HOWTO] for generic Linux instructions. |
| | 310 | |
| | 311 | ---- |
| | 312 | |
| 452 | | |
| 453 | | |
| 454 | | |
| 455 | | |
| 456 | | <!-- Selftest failed but status is 'PASS' --> |
| 457 | | <h3><a name="selftest-failed"></a>My ATA drive is failing its self-tests, but its SMART health status is 'PASS'. What's going on?</h3> |
| 458 | | |
| 459 | | <p>If your ATA drive supports self-tests, you should run them on a |
| 460 | | regular basis, for example one per week: |
| 461 | | <br/><span class="code">smartctl -t long /dev/hd?</span><br/> |
| 462 | | After the test has completed, you should examine the results with: |
| 463 | | <br/><span class="code">smartctl -l selftest /dev/hd?</span></p> |
| 464 | | |
| 465 | | <p>If the drive fails a self-test, but still has '<tt>PASS</tt>' SMART health |
| 466 | | status, this usually means that there is a corrupted (uncorrectable=UNC) sector on the |
| 467 | | disk. This means that the ECC data stored at that sector is not |
| 468 | | consistent with the user data stored at that sector, and an attempt to read the sector fails with a UNC error. |
| 469 | | This can be a one-time transient effect: a sudden power failure |
| 470 | | while the disk was writing to the sector corrupted the |
| 471 | | ECC code or data, but the sector <em>could</em> correctly store new data. |
| 472 | | Or it can be a permanent effect: the magnetic media |
| 473 | | has been damaged by a bit of dust, and the sector could <em>not</em> correctly store new data.</p> |
| 474 | | |
| 475 | | <p>If the disk can read the |
| 476 | | sector of data a single time, and the damage is permanent, not transient, then the disk firmware will mark the |
| 477 | | sector as 'bad' and allocate a spare sector to replace it. But |
| 478 | | if the disk can't read the sector even once, then it won't reallocate |
| 479 | | the sector, in hopes of being able, at some time in the future, to |
| 480 | | read the data from it. <b>A write to an unreadable (corrupted) sector will fix the problem.</b> |
| 481 | | If the damage is transient, then new consistent data will be written to the sector. |
| 482 | | If the damange is permanent, then the write will force sector reallocation. |
| 483 | | Please see <a |
| 484 | | href="badblockhowto.html" target="_blank">Bad block HOWTO</a> |
| 485 | | for instructions about how to force this sector to reallocate (Linux only).</p> |
| 486 | | |
| 487 | | <p>The disk still has passing health status because the firmware has not |
| 488 | | found other signs of trouble, such as a failing servo.</p> |
| 489 | | |
| 490 | | <p>Such disks can often be repaired by using the disk manufaturer's 'disk |
| 491 | | evaluation and repair' utility. Beware: this may force reallocation |
| 492 | | of the lost sector and thus corrupt or destroy any file system on the |
| 493 | | disk. See <a |
| 494 | | href="badblockhowto.html" target="_blank">Bad block HOWTO</a> |
| 495 | | for generic Linux instructions.</p> |
| 496 | | |
| 497 | | <hr /> |
